Rootzone reality: technical review on network performance
In July 2014 a network of passive-wick tension fluxmeters was established on 12 commercial cropping farms in the Canterbury, Manawatu, Hawke’s Bay, Waikato and Auckland regions. The network is being used to quantify leaching losses of nitrogen (N) and phosphorus (P) in drainage water from below the crop rootzone. In this report we review the technical performance of the fluxmeter network for the period 1 October 2014 to 28 February 2018. This includes an
evaluation of measured drainage volumes against modelled outputs (generated using a mechanistic soil water balance model) and a summary of N and P losses in drainage water from each site. At most sites, this reporting period represents at least 3 years of continuous data collection allowing for multiple crop rotations and the dissipation of settling effects following installation of the fluxmeter units.
From this review we conclude that eleven of the twelve fluxmeter sites are operating in a manner suitable for the generation of robust data on N and P losses in drainage water (Table i). Only at one site (Site 8) were serious concerns raised about the performance of some fluxmeters, a number of which captured no drainage in the 41 months of monitoring. At all other sites, the fluxmeter units were found to be operational and effective at capturing drainage,
despite sometimes large variations in volumes captured by individual units. Drainage variability was attributed to different patterns of accumulation across the sites as affected by topography, soil physical properties and crop factors. At seven of the sites (Sites 1, 2, 3, 4, 5, 6 and 10), captured drainage volumes were highly consistent with the timing of drainage events and patterns of drainage accumulation as predicted by the soil water balance model (Table i). At the remaining five sites (excluding Site 8), and during certain periods, captured drainage volumes deviated from modelled predictions. This was attributed to flooding of the fluxmeters resulting in excess capture volumes (Sites 7, 11 and 12) or inefficient drainage collection resulting in reduced capture volumes (Site 9). Where flooding occurred, N and P losses were estimated using a combination of modelled drainage and measured concentration data, an approach that is common for other sampling methods such as suction cups. Concentration data from flooded units were carefully evaluated to ensure that values were representative of drainage permeating through the soil profile, as opposed to bypass flow or ground water infiltration. At Site 9, we are awaiting soil physical characterisation results to confirm whether captured volumes are indeed being underestimated.
Inorganic N losses have varied widely across the network sites ranging from 0.5 to 234 kg N/ha (Year 1; 1 September 2014 to 30 August 2015), 2 to 173 kg N/ha (Year 2; 1 September 2015 to 30 August 2016), 16 to 203 kg N/ha (Year 3; 1 September 2016 to 30 August 2017) and 3 to 118 kg N/ha (Year 4, part; 1 September 2017 to 28 February 2018) (Figure i). High net losses were associated with high drainage volumes and inorganic N concentrations (predominantly
nitrate-N) in drainage water. In most cases drainage losses occurred during the late autumn, winter and/or early spring months when rainfall and soil moisture contents were highest. Cumulative P losses in the respective Year 1, Year 2, Year 3 and current Year 4 monitoring periods ranged from 0.02 to 1.99 kg P/ha, 0.05 to 0.28 kg P/ha, 0.04 to 0.67 kg P/ha and 0.01 to 0.10 kg P/ha (Figure i). These represented fairly small net losses, of which the majority (50–95%) was in the dissolved reactive form (DRP) in the drainage water.